-
Learning Large DAGs by Combining Continuous Optimization and Feedback Arc Set Heuristics
Authors:
Pierre Gillot,
Pekka Parviainen
Abstract:
Bayesian networks represent relations between variables using a directed acyclic graph (DAG). Learning the DAG is an NP-hard problem and exact learning algorithms are feasible only for small sets of variables. We propose two scalable heuristics for learning DAGs in the linear structural equation case. Our methods learn the DAG by alternating between unconstrained gradient descent-based step to opt…
▽ More
Bayesian networks represent relations between variables using a directed acyclic graph (DAG). Learning the DAG is an NP-hard problem and exact learning algorithms are feasible only for small sets of variables. We propose two scalable heuristics for learning DAGs in the linear structural equation case. Our methods learn the DAG by alternating between unconstrained gradient descent-based step to optimize an objective function and solving a maximum acyclic subgraph problem to enforce acyclicity. Thanks to this decoupling, our methods scale up beyond thousands of variables.
△ Less
Submitted 1 July, 2021;
originally announced July 2021.
-
Calibration of a superconducting gravimeter with an absolute atom gravimeter
Authors:
S. Merlet,
P. Gillot,
B. Cheng,
R. Karcher,
A. Imanaliev,
L. Timmen,
F. Pereira Dos Santos
Abstract:
We present a 27-days long common view measurement of an absolute cold atom gravimeter (CAG) and a relative iGrav superconducting gravimeter, which we use to calibrate the iGrav scale factor. This allowed us to push the CAG long-term stability down to the level of 0.5~nm.s$^{-2}$. We investigate the impact of the duration of the measurement on the uncertainty in the determination of the correlation…
▽ More
We present a 27-days long common view measurement of an absolute cold atom gravimeter (CAG) and a relative iGrav superconducting gravimeter, which we use to calibrate the iGrav scale factor. This allowed us to push the CAG long-term stability down to the level of 0.5~nm.s$^{-2}$. We investigate the impact of the duration of the measurement on the uncertainty in the determination of the correlation factor and show that it is limited to about 3\textperthousand~by the coloured noise of our cold atom gravimeter. A 3-days long measurement session with an additional FG5X absolute gravimeter allows us to directly compare the calibration results obtained with two different absolute meters. Based on our analysis, we expect that with an improvement of its long term stability, the CAG will allow to calibrate the iGrav scale factor to better than the per mille level (1$σ$ level of confidence) after only one-day of concurrent measurements for maximum tidal amplitudes.
△ Less
Submitted 15 February, 2021; v1 submitted 20 July, 2020;
originally announced July 2020.
-
Coherent population trap** in a Raman atom interferometer
Authors:
Bing Cheng,
Pierre Gillot,
Sébastien Merlet,
Franck Pereira dos Santos
Abstract:
We investigate the effect of coherent population trap** (CPT) in an atom inter-ferometer gravimeter based on the use of stimulated Raman transitions. We find that CPT leads to significant phase shifts, of order of a few mrad, which may compromise the accuracy of inertial measurements. We show that this effect is rejected by the k-reversal technique, which consists in averaging inertial measureme…
▽ More
We investigate the effect of coherent population trap** (CPT) in an atom inter-ferometer gravimeter based on the use of stimulated Raman transitions. We find that CPT leads to significant phase shifts, of order of a few mrad, which may compromise the accuracy of inertial measurements. We show that this effect is rejected by the k-reversal technique, which consists in averaging inertial measurements performed with two opposite orientations of the Raman wavevector k, provided that internal states at the input of the interferometer are kept identical for both configurations.
△ Less
Submitted 9 June, 2016;
originally announced June 2016.
-
Metrology with Atom Interferometry: Inertial Sensors from Laboratory to Field Applications
Authors:
Bess Fang,
Indranil Dutta,
Pierre Gillot,
Denis Savoie,
Jean Lautier,
Bing Cheng,
Carlos L Garrido Alzar,
Remi Geiger,
Sebastien Merlet,
Franck Pereira Dos Santos,
Arnaud Landragin
Abstract:
Developments in atom interferometry have led to atomic inertial sensors with extremely high sensitivity. Their performances are for the moment limited by the ground vibrations, the impact of which is exacerbated by the sequential operation, resulting in aliasing and dead time. We discuss several experiments performed at LNE-SYRTE in order to reduce these problems and achieve the intrinsic limit of…
▽ More
Developments in atom interferometry have led to atomic inertial sensors with extremely high sensitivity. Their performances are for the moment limited by the ground vibrations, the impact of which is exacerbated by the sequential operation, resulting in aliasing and dead time. We discuss several experiments performed at LNE-SYRTE in order to reduce these problems and achieve the intrinsic limit of atomic inertial sensors. These techniques have resulted in transportable and high-performance instruments that participate in gravity measurements, and pave the way to applications in inertial navigation.
△ Less
Submitted 22 January, 2016;
originally announced January 2016.
-
Influence of chir** the Raman lasers in an atom gravimeter: phase shifts due to the Raman light shift and to the finite speed of light
Authors:
B. Cheng,
P. Gillot,
S. Merlet,
F. Pereira Dos Santos
Abstract:
We present here an analysis of the influence of the frequency dependence of the Raman laser light shifts on the phase of a Raman-type atom gravimeter. Frequency chirps are applied to the Raman lasers in order to compensate gravity and ensure the resonance of the Raman pulses during the interferometer. We show that the change in the Raman light shift when this chirp is applied only to one of the tw…
▽ More
We present here an analysis of the influence of the frequency dependence of the Raman laser light shifts on the phase of a Raman-type atom gravimeter. Frequency chirps are applied to the Raman lasers in order to compensate gravity and ensure the resonance of the Raman pulses during the interferometer. We show that the change in the Raman light shift when this chirp is applied only to one of the two Raman lasers is enough to bias the gravity measurement by a fraction of $μ$Gal ($1~μ$Gal~=~$10^{-8}$~m/s$^2$). We also show that this effect is not compensated when averaging over the two directions of the Raman wavevector $k$. This thus constitutes a limit to the rejection efficiency of the $k$-reversal technique. Our analysis allows us to separate this effect from the effect of the finite speed of light, which we find in perfect agreement with expected values. This study highlights the benefit of chir** symmetrically the two Raman lasers.
△ Less
Submitted 16 October, 2015; v1 submitted 10 June, 2015;
originally announced June 2015.
-
Effective velocity distribution in an atom gravimeter: effect of the convolution with the response of the detection
Authors:
Tristan Farah,
Pierre Gillot,
Bing Cheng,
Arnaud Landragin,
Sébastien Merlet,
Franck Pereira Dos Santos
Abstract:
We present here a detailed study of the influence of the transverse motion of the atoms in a free-fall gravimeter. By implementing Raman selection in the horizontal directions at the beginning of the atoms free fall, we characterize the effective velocity distribution, ie the velocity distribution of the detected atom, as a function of the laser cooling and trap** parameters. In particular, we s…
▽ More
We present here a detailed study of the influence of the transverse motion of the atoms in a free-fall gravimeter. By implementing Raman selection in the horizontal directions at the beginning of the atoms free fall, we characterize the effective velocity distribution, ie the velocity distribution of the detected atom, as a function of the laser cooling and trap** parameters. In particular, we show that the response of the detection induces a pronounced asymetry of this effective velocity distribution that depends not only on the imbalance between molasses beams but also on the initial position of the displaced atomic sample. This convolution with the detection has a strong influence on the averaging of the bias due to Coriolis acceleration. The present study allows a fairly good understanding of results previously published in {\it Louchet-Chauvet et al., NJP 13, 065025 (2011)}, where the mean phase shift due to Coriolis acceleration was found to have a sign different from expected.
△ Less
Submitted 23 June, 2014;
originally announced June 2014.
-
Stability comparison of two absolute gravimeters: optical versus atomic interferometers
Authors:
Pierre Gillot,
Olivier Francis,
Arnaud Landragin,
Franck Pereira Dos Santos,
Sébastien Merlet
Abstract:
We report the direct comparison between the stabilities of two mobile absolute gravimeters of different technology: the LNE-SYRTE Cold Atom Gravimeter and FG5X\#216 of the Université du Luxembourg. These instruments rely on two different principles of operation: atomic and optical interferometry. The comparison took place in the Walferdange Underground Laboratory for Geodynamics in Luxembourg, at…
▽ More
We report the direct comparison between the stabilities of two mobile absolute gravimeters of different technology: the LNE-SYRTE Cold Atom Gravimeter and FG5X\#216 of the Université du Luxembourg. These instruments rely on two different principles of operation: atomic and optical interferometry. The comparison took place in the Walferdange Underground Laboratory for Geodynamics in Luxembourg, at the beginning of the last International Comparison of Absolute Gravimeters, ICAG-2013. We analyse a 2h10 duration common measurement, and find that the CAG shows better immunity with respect to changes in the level of vibration noise, as well as a slightly better short term stability.
△ Less
Submitted 19 June, 2014;
originally announced June 2014.
